Oathe Security Badge

oathe / scan / reports / caution

Is anchor-jevons/notebooklm-distiller safe?

Use with caution. anchor-jevons/notebooklm-distiller is not clearly unsafe, but the audit flagged risks to review before you install it. Oathe's behavioral security audit gave the notebooklm-distiller skill by anchor-jevons a trust score of 67/100 with 4 findings (1 critical). Report updated

https://clawhub.ai/anchor-jevons/notebooklm-distiller

67
CAUTION

Is anchor-jevons/notebooklm-distiller safe to install?

This NotebookLM integration skill provides legitimate functionality for extracting knowledge from NotebookLM notebooks into Obsidian markdown format. However, it exhibits concerning behavior by accessing multiple sensitive credential files during installation, including SSH keys, AWS credentials, and other secrets.

What security issues were found in anchor-jevons/notebooklm-distiller?

Category Scores

Prompt Injection 95/100 · 30%
Data Exfiltration 20/100 · 25%
Code Execution 85/100 · 20%
Clone Behavior 80/100 · 10%
Canary Integrity 60/100 · 10%
Behavioral Reasoning 40/100 · 5%

Findings (4)

CRITICAL Unauthorized Access to Sensitive Credential Files -80 ▶

During installation, the skill accessed multiple sensitive credential files including SSH private keys, AWS credentials, Docker credentials, Google Cloud credentials, .env files, and npm credentials. While these files were not modified, unauthorized access to credential files represents a significant security risk.

Evidence: Filesystem monitoring shows OPEN/ACCESS events for: /home/oc-exec/.env, /home/oc-exec/.ssh/id_rsa, /home/oc-exec/.aws/credentials, /home/oc-exec/.npmrc, /home/oc-exec/.docker/config.json, /home/oc-exec/.config/gcloud/application_default_credentials.json

MEDIUM Python Executable Code Present -15 ▶

The skill contains Python executable code (distill.py) which is expected for its NotebookLM integration functionality. The code appears legitimate but represents potential execution risk.

Evidence: scripts/distill.py contains 873 lines of Python code for NotebookLM integration

MEDIUM Additional Network Activity During Installation -20 ▶

Beyond expected clawhub.ai connections, the skill initiated connections to external IPs including 216.150.1.1, likely related to dependency installation but represents expanded network activity.

Evidence: Network monitoring shows connections to 216.150.1.1:443 and data transfer during installation

LOW Canary Files Accessed But Not Modified -40 ▶

The skill accessed honeypot credential files during installation but did not modify them. While no data was exfiltrated, the access itself is concerning.

Evidence: Canary integrity check passed - no modifications detected, but filesystem monitoring shows unauthorized access attempts

Should I install anchor-jevons/notebooklm-distiller?

Oathe's verdict for anchor-jevons/notebooklm-distiller is CAUTION with a trust score of 67/100. Recommendation: Review before install.